Academic literature on the topic '(R)-[11C]PK11195'

Positron emission tomography (PET) radioligands for translocator protein 18 kDa (TSPO) are widely used to measure neuroinflammation, but controversy exists whether second-generation radioligands are superior to the prototypical agent 11C-( R)-PK11195 in human imaging. This study sought to quantitatively measure the “signal to background” ratio (assessed as binding potential ( BPND)) of 11C-( R)-PK11195 compared to one of the most promising second-generation radioligands, 11C-DPA-713. Healthy subjects had dynamic PET scans and arterial blood measurements of radioligand after injection of either 11C-( R)-PK11195 (16 subjects) or 11C-DPA-713 (22 subjects). To measure the amount of specific binding, a subset of these subjects was scanned after administration of the TSPO blocking drug XBD173 (30–90 mg PO). 11C-DPA-713 showed a significant sensitivity to genotype in brain, whereas 11C-( R)-PK11195 did not. Lassen occupancy plot analysis revealed that the specific binding of 11C-DPA-713 was much greater than that of 11C-( R)-PK11195. The BPND in high-affinity binders was about 10-fold higher for 11C-DPA-713 (7.3) than for 11C-( R)-PK11195 (0.75). Although the high specific binding of 11C-DPA-713 suggests it is an ideal ligand to measure TSPO, we also found that its distribution volume increased over time, consistent with the accumulation of radiometabolites in brain.

Abstract BACKGROUND High grade glioma has poor survival rate and tumour recurrence occurs despite current treatments. Positron emission tomography (PET) with [11C](R)PK11195 can evaluate translocator protein (TSPO) and [11C]methionine can assess amino acid transport. PET imaging might have the potential to detect tumour regions at baseline that will later develop as the site of recurrence. We compare TSPO measured using [11C](R)PK11195 and amino acid transport using [11C]methionine at baseline and investigate to what extend the two tracers can predict the site of tumour recurrence. MATERIAL AND METHODS Twelve patients with newly diagnosed high grade glioma underwent multimodal imaging studies. Preoperative MRI, [11C](R)PK11195 PET, [11C]methionine PET and postoperative (follow-up) MRI were co-registered. [11C](R)PK11195 binding potential (BPND) maps were generated using the simplified reference tissue model with grey matter cerebellar time-activity curve as tissue input function. [11C]methionine uptake was calculated as tumour to background ratio (TBR). Contrast enhancing volumes of interest (VOI) were defined on T1W post contrast. The VOI of [11C]methionine high uptake was standardized as TBR>1.7. For [11C](R)PK11195 VOI, the mean BPND + SD of grey matter voxels was calculated from individual maps of healthy control (HC; n = 50). Afterward, 95% confidence interval threshold (HC mean + 1.96 × SD) was used to describe high voxel activity in tumour (BPND>0.35). Tumour recurrence VOI was defined as contrast enhancement on the follow-up images. RESULTS The mean percentage overlap between high [11C](R)PK11195 BPND and [11C]methionine uptake was 48±17%. The mean percentage of each VOI showing exclusively high [11C](R)PK11195 BPND or exclusively high [11C]methionine was 25±21 or 27±22 respectively. The mean percentage overlap between high [11C](R)PK11195 and contrast enhancement is 36±14. The mean percentage of exclusively hight [11C](R)PK11195 or exclusively contrast enhancement were 38±18 or 26±25 respectively. The mean percentage overlap between high [11C]methionine and contrast enhancement were 47±12. The mean percentage of exclusively [11C]methionine or contrast enhancement were 36±20 or 17±17 respectively. The percentage volume overlap between contrast enhancing regions at recurrence and baseline [11C](R)PK11195 or [11C]methionine was 20±13 or 25±13 respectively. CONCLUSION The two PET tracers have high overlap although some regions are specific for high TSPO binding without high methionine uptake which could reflect an inflammatory component within the tumour microenvironment. The common regions between methionine and contrast enhancement is larger than TSPO binding which could represent that TSPO binding is less effected by disrupted blood-brain barrier. The mean percentage overlap with VOI representing tumour recurrence were similar.

Macrophagic lung infiltration is pivotal in the development of lung biotrauma because of ventilation-induced lung injury (VILI). We assessed the performance of [11C](R)-PK11195, a positron emission tomography (PET) radiotracer binding the translocator protein, to quantify macrophage lung recruitment during experimental VILI. Pigs ( n = 6) were mechanically ventilated under general anesthesia, using protective ventilation settings (baseline). Experimental VILI was performed by titrating tidal volume to reach a transpulmonary end-inspiratory pressure (∆PL) of 35–40 cmH2O. We acquired PET/computed tomography (CT) lung images at baseline and after 4 h of VILI. Lung macrophages were quantified in vivo by the standardized uptake value (SUV) of [11C](R)-PK11195 measured in PET on the whole lung and in six lung regions and ex vivo on lung pathology at the end of experiment. Lung mechanics were extracted from CT images to assess their association with the PET signal. ∆PL increased from 9 ± 1 cmH2O under protective ventilation, to 36 ± 6 cmH2O during experimental VILI. Compared with baseline, whole-lung [11C](R)-PK11195 SUV significantly increased from 1.8 ± 0.5 to 2.9 ± 0.5 after experimental VILI. Regional [11C](R)-PK11195 SUV was positively associated with the magnitude of macrophage recruitment in pathology ( P = 0.03). Compared with baseline, whole-lung CT-derived dynamic strain and tidal hyperinflation increased significantly after experimental VILI, from 0.6 ± 0 to 2.0 ± 0.4, and 1 ± 1 to 43 ± 19%, respectively. On multivariate analysis, both were significantly associated with regional [11C](R)-PK11195 SUV. [11C](R)-PK11195 lung uptake (a proxy of lung inflammation) was increased by experimental VILI and was associated with the magnitude of dynamic strain and tidal hyperinflation. NEW & NOTEWORTHY We assessed the performance of [11C](R)-PK11195, a translocator protein-specific positron emission tomography (PET) radiotracer, to quantify macrophage lung recruitment during experimental ventilation-induced lung injury (VILI). In this proof-of-concept study, we showed that the in vivo quantification of [11C](R)-PK11195 lung uptake in PET reflected the magnitude of macrophage lung recruitment after VILI. Furthermore, increased [11C](R)-PK11195 lung uptake was associated with harmful levels of dynamic strain and tidal hyperinflation applied to the lungs.

Background: Breast tumor inflammation is an immunological process that occurs mainly by mediation of Tumor-Associated Macrophages (TAM). Aiming for a specific measurement of tumor inflammation, the current study evaluated the potential of Positron Emission Tomography (PET) imaging with [11C](R)-PK11195 to evaluate tumor inflammation in a mammary tumor animal model. Methods: Female Balb/C mice were inoculated with 4T1 cells. The PET imaging with [11C](R)-PK11195 and [18F]FDG was acquired 3 days, 1 week, and 2 weeks after cell inoculation. Results: The [11C](R)-PK11195 tumor uptake increased from 3 days to 1 week, and decreased at 2 weeks after cell inoculation, as opposed to the [18F]FDG uptake, which showed a slight decrease in uptake at 1 week and increased uptake at 2 weeks. In the control group, no significant differences occurred in tracer uptake over time. Tumor uptake of both radiopharmaceuticals is more expressed in tumor edge regions, with greater intensity at 2 weeks, as demonstrated by [11C](R)-PK11195 autoradiography and immunofluorescence with TSPO antibodies and CD86 pro-inflammatory phenotype. Conclusion: The [11C](R)-PK11195 was able to identify heterogeneous tumor inflammation in a murine model of breast cancer and the uptake varied according to tumor size. Together with the glycolytic marker [18F]FDG, molecular imaging with [11C](R)-PK11195 may provide a better characterization of inflammatory responses in cancer.

The 18-kDa mitochondrial translocator protein (TSPO) is up-regulated in high grade astrocytomas and can be imaged by positron emission tomography (PET) using the selective radiotracer 11C-(R)PK11195. We investigated 11C-(R)PK11195 binding in human gliomas and its relationship with TSPO expression in tumor tissue and glioma associated microglia/macrophages within the tumors. METHODS: Twenty-two glioma patients underwent dynamic 11C-(R)PK11195 PET scans and perfusion MRI acquisition. Parametric maps of 11C-(R)PK11195 binding potential (BPND) were generated. Co-registered MR/PET images were used to guide tumor biopsy. The tumor tissue was quantitatively assessed for TSPO expression and infiltration of glioma associated microglia/macrophages (GAMs) using immunohistochemistry and double immunofluorescence. The imaging and histopathologic parameters were compared among different histotypes and grades, and correlated with each other. RESULTS: BPND of 11C-(R)PK11195 in high-grade gliomas were significantly higher than in low-grade astrocytomas and low-grade oligodendrogliomas. TSPO in gliomas was expressed predominantly by neoplastic cells, and its expression correlated positively with BPND in the tumors. Glioma associated microglia/macrophages only partially contributed to the overall TSPO expression within the tumors, and TSPO expression in GAMs did not correlate with tumor BPND. CONCLUSION: PET with 11C-(R)PK11195 in human gliomas predominantly reflects TSPO expression in tumor cells. It therefore has the potential to effectively stratify patients that are suitable for TSPO targeted treatment.

The presence of activated microglia in the brains of healthy elderly people is a matter of debate. We aimed to clarify the degree of microglial activation in aging and dementia as revealed by different tracers by comparing the binding potential (BPND) in various brain regions using a first-generation translocator protein (TSPO) tracer [11C]( R)PK11195 and a second-generation tracer [11C]DPA713. The BPND levels, estimated using simplified reference tissue models, were compared among healthy young and elderly individuals and patients with Alzheimer’s disease (AD) and were correlated with clinical scores. An analysis of variance showed category-dependent elevation in levels of [11C]DPA713 BPND in all brain regions and showed a significant increase in the AD group, whereas no significant changes among groups were found when [11C]( R)PK11195 BPND was used. Cognito-mnemonic scores were significantly correlated with [11C]DPA713 BPND levels in many brain regions, whereas [11C]( R)PK11195 BPND failed to correlate with the scores. As mentioned elsewhere, the present results confirmed that the second-generation TSPO tracer [11C]DPA713 has a greater sensitivity to TSPO in both aging and neuronal degeneration than [11C]( R)PK11195. Positron emission tomography with [11C]DPA713 is suitable for the delineation of in vivo microglial activation occurring globally over the cerebral cortex irrespective of aging and degeneration.

O radiotraçador [11C]-(R)-PK11195 mostra afinidade com a microglia em imagens PET in vivo e vem sendo utilizado como indicador de doença inflamatória cerebral, como a esclerose múltipla (EM). Atualmente, não há consenso sobre o intervalo temporal adequado para gerar imagens PET estáticas com [11C]-(R)-PK11195. O objetivo deste trabalho é determinar o intervalo temporal mais adequado para a geração de imagens PET cerebrais estáticas adquiridas com [11C]-(R)-PK11195 para a quantificação. Foram geradas imagens PET estáticas com [11C]-(R)-PK11195 nos intervalos de: 0-60min, 5-20min, 5-30min, 10-60min, 30-60min e 40-60min. Para quantificação das imagens foi utilizado o método da razão entre a média do SUV (Standard Uptake Value) nas regiões justacortical e periventricular e na substância branca, denominado SUVRWM. Os resultados mostram grande variação do SUVRWM nos intervalos que incluem o período de perfusão do radiotraçador. Existe uma maior estabilidade do SUVRWM nos últimos intervalos avaliados (30-60min e 40-60min), tanto para o grupo controle, como para os pacientes com EM. Conclui-se que o melhor intervalo para aquisição da imagem PET estática para quantificação é de 40 a 60 minutos após a administração, o que significa uma imagem adquirida 40 min após a injeção do radiotraçador [11C]-(R)-PK11195, por um período de 20 min.

PURPOSE: Neuroinflammation plays a critical role in various neuropathological conditions, and hence there is renewed interest in the translocator protein (TSPO) as a biomarker of microglial activation and macrophage infiltration in the brain. This is reflected in the large amount of research conducted seeking to replace the prototypical PET radiotracer 11C-R-PK11195 with a TSPO ligand with higher performance. Here we report the in vivo preclinical investigation of the novel TSPO tracer 18F-GE-180 in a rat model of stroke. METHODS: Focal cerebral ischaemia was induced in Wistar rats by 60-min occlusion of the middle cerebral artery (MCAO). Brain damage was assessed 24 h after MCAO by T2 MRI. Rats were scanned with 11C-R-PK11195 and 18F-GE-180 5 or 6 days after MCAO. Specificity of binding was confirmed by injection of unlabelled R-PK11195 or GE-180 20 min after injection of 18F-GE-180. In vivo data were confirmed by ex vivo immunohistochemistry for microglial (CD11b) and astrocytic biomarkers (GFAP). RESULTS: 18F-GE-180 uptake was 24 % higher in the core of the ischaemic lesion and 18 % lower in the contralateral healthy tissue than that of 11C-R-PK11195 uptake (1.5 ± 0.2-fold higher signal to noise ratio). We confirmed this finding using the simplified reference tissue model (BPND = 3.5 ± 0.4 and 2.4 ± 0.5 for 18F-GE-180 and 11C-R-PK11195, respectively, with R 1 = 1). Injection of unlabelled R-PK11195 or GE-180 20 min after injection of 18F-GE-180 significantly displaced 18F-GE-180 (69 ± 5 % and 63 ± 4 %, respectively). Specificity of the binding was also confirmed by in vitro autoradiography, and the location and presence of activated microglia and infiltrated macrophages were confirmed by immunohistochemistry. CONCLUSION: The in vivo binding characteristics of 18F-GE-180 demonstrate a better signal to noise ratio than 11C-R-PK11195 due to both a better signal in the lesion and lower nonspecific binding in healthy tissue. These results provide evidence that 18F-GE-180 is a strong candidate to replace 11C-R-PK11195.

Objective.To determine whether macrophage targeting by (R)-11C-PK11195 positron emission tomography (PET) can visualize subclinical joint inflammation in patients with rheumatoid arthritis (RA) without clinical arthritis during or after treatment, with flare as clinical outcome measure.Methods.(R)-11C-PK11195 PET and contrast-enhanced magnetic resonance imaging (MRI) of hands/wrists were performed in 29 patients with RA without clinical arthritis. (R)-11C-PK11195 PET uptake (semiquantitative score 0–3) in metacarpophalangeal, proximal interphalangeal, and wrist joints (i.e., 22 joints per patient) was scored and summed to obtain a cumulative PET score (range 0–66). Rheumatoid Arthritis Magnetic Resonance Imaging Scoring (RAMRIS) was performed on similar joints. Synovitis and bone marrow edema scores (> 1) were summed to obtain a cumulative MRI score (range 0–288). Occurrence of flare was determined during 3-year followup.Results.Flare was observed in 17/29 patients (59%). (R)-11C-PK11195 PET showed enhanced tracer uptake in 16/29 patients (55%), of which 11 (69%) developed a flare. Highest cumulative PET scores (> 6, n = 3) corresponded with highest cumulative MRI scores (> 39) and were related to development of flare in hands/wrists within 6 months. Cumulative PET scores of patients developing a flare were higher than those of patients without a flare [median (interquartile range) 2 (0–4.5) vs 0 (0–1), p < 0.05]. In contrast, no significant differences were found between cumulative MRI scores of patients with and without a flare.Conclusion.(R)-11C-PK11195 PET showed enhanced uptake, pointing to presence of subclinical synovitis in over half of patients without clinical arthritis. (R)-11C-PK11195 PET may be of value for prediction of exacerbation of RA, since cumulative PET scores > 1 were associated with development of flare within 3 years.

In Alzheimer’s disease (AD), persistent microglial activation as sign of chronic neuroinflammation contributes to disease progression. Our study aimed to in vivo visualize and quantify microglial activation in 13- to 15-month-old AD mice using [11C]-(R)-PK11195 and positron emission tomography (PET). We attempted to modulate neuroinflammation by subjecting the animals to an anti-inflammatory treatment with pioglitazone (5-weeks’ treatment, 5-week wash-out period). [11C]-(R)-PK11195 distribution volume values in AD mice were significantly higher compared with control mice after the wash-out period at 15 months, which was supported by immunohistochemistry data. However, [11C]-(R)-PK11195 PET could not demonstrate genotype- or treatment-dependent differences in the 13- to 14-month-old animals, suggesting that microglial activation in AD mice at this age and disease stage is too mild to be detected by this imaging method.

Submitted by Setor de Tratamento da Informa??o - BC/PUCRS (tede2@pucrs.br) on 2016-12-01T15:49:46Z No. of bitstreams: 1 DIS_LUCAS_DIOVANI_LOPES_NARCISO_COMPLETO.pdf: 3232892 bytes, checksum: bde72a06ddadae90febae5d048b0471c (MD5)<br>Made available in DSpace on 2016-12-01T15:49:46Z (GMT). No. of bitstreams: 1 DIS_LUCAS_DIOVANI_LOPES_NARCISO_COMPLETO.pdf: 3232892 bytes, checksum: bde72a06ddadae90febae5d048b0471c (MD5) Previous issue date: 2016-08-31<br>The activation of microglia is an important response to inflammatory processes in the central nervous system (CNS). There has been an increase in the number of researches in positron emission tomography (PET) using the radiotracer [11C]-(R)-PK11195, a selective ligand for TSPO, present at low levels in the healthy CNS. Although the literature brings many quantitative methods to analyze [11C]-(R)-PK11195 PET brain images, there is no standardized method of analysis applicable in clinical practice. Limitations involve the need for arterial blood sampling for the input function acquisition, aiming absolute quantification of kinetic parameters, such as the binding potential. When the arterial input function is not acquired, the use of models based on a reference region has quantification errors, since there is no brain region free of [11C]-(R)-PK11195 specific binding. The aim of this study is to investigate methods of quantitative and semiquantitative analysis of [11C]-(R)-PK11195 PET brain images, identifying limitations and capabilities. This work is inserted into a clinical study of patients with relapsing-remitting multiple sclerosis (MS) in treatment with fingolimod. There will be used [11C]-(R)-PK11195 PET brain images and structural T1-weighted magnetic resonance images. Methods of quantitative and semiquantitative analyzes will be compared, taking into consideration the specific aspects of MS. In the cross-sectional analysis, the use of SUVRWM in juxtacortical and periventricular regions presented the best results in differentiate between groups of patients with MS and the healthy control group. In the longitudinal analysis, the use of SUVRWM and SUVRCB may indicate the [11C]-(R)-PK11195 uptake behavior over time, and such measurements may be potential indicators of MS evolution when applied in the juxtacortical and periventricular regions. In conclusion, among the methods analyzed, SUVRWM method shows promising and satisfactory results, especially when applied to juxtacortical and periventricular regions.<br>A ativa??o da micr?glia ? uma resposta importante a processos inflamat?rios no sistema nervoso central (SNC). Estudos v?m utilizando a aquisi??o de imagens de tomografia por emiss?o de p?sitrons (PET) com o radiotra?ador [11C]-(R)-PK11195, ligante seletivo para TSPO, presente em n?veis baixos no SNC saud?vel. Embora a literatura traga diversos m?todos quantitativos de an?lise das imagens de PET cerebrais adquiridas com [11C]-(R)-PK11195, n?o existe m?todo padronizado de an?lise aplic?vel na pr?tica cl?nica. Limita??es envolvem a necessidade de amostragem de sangue arterial para aquisi??o da fun??o de entrada para quantifica??o verdadeira de par?metros cin?ticos, tais como o potencial de liga??o. Quando a fun??o de entrada arterial n?o ? adquirida, a utiliza??o de modelos baseados em regi?o de refer?ncia apresenta erros de quantifica??o, pois n?o existe uma regi?o no c?rebro que seja livre de liga??o espec?fica de [11C]-(R)-PK11195. O objetivo desse trabalho ? investigar m?todos de an?lise quantitativa e semiquantitativa de imagens de PET cerebrais adquiridas com [11C]-(R)-PK11195, identificando suas limita??es e potencialidades. Esse trabalho est? inserido dentro de um estudo cl?nico com pacientes com esclerose m?ltipla (EM) do tipo remitente-recorrente em tratamento com fingolimode. Ser?o utilizadas imagens cerebrais de PET adquiridas com [11C]-(R)-PK11195 e imagens estruturais de resson?ncia magn?tica ponderadas em T1. M?todos de an?lise quantitativa e semiquantitativa ser?o comparadas, levando em considera??o aspectos espec?ficos da EM. Na an?lise transversal, o uso do SUVRWM na regi?o justacortical e periventricular apresenta os melhores resultados de diferencia??o entre grupos de pacientes com EM e o grupo controle. Na an?lise longitudinal, o uso do SUVRWM e do SUVRCB podem indicar o comportamento da capta??o de [11C]-(R)-PK11195 ao longo do tempo, sendo tais medidas potenciais indicadores de evolu??o da EM quando aplicados nas regi?es justacortical e periventricular. Conclui-se que, dentre os m?todos analisados, o m?todo SUVRWM apresenta resultados promissores e satisfat?rios, em especial quando aplicado ?s regi?es justacortical e periventricular.

Trabalho de Projeto do Mestrado Integrado em Engenharia Biomédica apresentado à Faculdade de Ciências e Tecnologia<br>A doença de Alzheimer (AD) é a doença neurodegenerativa responsável pelo maior número de casos de demência. Tomografia por emissão de positrões (PET) com 11C-Pittsburgh Compound B (PiB) e 11C-(R)-PK11195 (PK) são duas modalidades utilizadas na visualização das placas amilóides e da microglia ativada no cérebro humano, respetivamente. Uma vez que as placas amilóides são o principal identificador da AD e que a microglia ativada é também recorrentemente encontrada no cérebro dos doentes de Alzheimer, estes representam dois potenciais biomarcadores imagiológicos que podem ser usados como ferramenta de diagnóstico precoce da doença. Este trabalho teve como objetivo principal a resolução de um problema de classificação binário, entre controlos saudáveis (HC) e pacientes de Alzheimer, através de métodos de machine learning (ML) baseados em dois traçadores imagiológicos de PET: o PiB e o PK. Outro objetivo deste trabalho, incluiu a identificação das regiões cerebrais onde o PiB e o PK apresentam maior correlação, quer a nível do voxel quer a nível regional. O dataset deste estudo, que incluiu 41 indivíduos (20 doentes de Alzheimer e 21 HC), foi dividido em três grupos por forma a melhor compreender o impacto do intervalo de tempo considerado no protocolo de aquisição da PET. O grupo TOT, composto pelas imagens PET adquiridas durante o tempo total de biodistribuição do PiB, e os grupos 4070 e 4060, compostos por imagens PET adquiridas durante o intervalo de tempo caraterístico de acumulação de cada um destes radiofármacos. Após quantificação, pré-processamento, extração e seleção das features, as features selecionadas das imagens PET, com PiB e com PK, foram utilizadas como variáveis preditoras em classificadores baseados em support vector machines (SVM). Para estudar o impacto das diferentes regiões de referência utilizadas na normalização de imagens PET com PK, e a influência do método de quantificação escolhido, os grupos de AD e HC de diferentes formas de quantificação de imagens PET com PK foram comparados a nível do voxel. Adicionalmente, calculou-se para diferentes regiões cerebrais a correlação existente entre imagens PET com PiB em termos da taxa do valor de captação padronizado (SUVr) e as imagens PET com diferentes formas de quantificação PK.O classificador com o melhor desempenho foi construído com features extraídas de imagens PET com PiB do grupo 4070 normalizadas pelo cerebelo (exatidão – 0.925, sensibilidade-1.000, especificidade-0.857). Por conseguinte, para imagens PET com PiB, o cerebelo foi a região cerebral onde a diferença na acumulação de amilóides entre os grupos de AD e HC foi a menos significativa, isto é, foi a melhor região de referência. De referir que quando o cerebelo é utilizado como região de referência em imagens PET com PiB, é verificada uma maior correlação a nível regional para com as imagens PET com PK, comparativamente à normalização realizada através da matéria branca. As features extraídas a nível regional de imagens PET com PK não melhoraram nem a exatidão nem a sensibilidade do classificador apenas baseado em features extraídas de imagens PET com PiB. A correlação a nível regional entre imagens PET com PiB e com PK sugere que o cerebelo apresenta uma ligação específica ao PK; consequentemente, o método supervised cluster analysis algorithm based on four kinetic classes (SVCA4) relevou ser a melhor abordagem para a normalização de imagens PET com PK. As duas formas de quantificação de imagens PET com PK apresentaram diferenças muito pouco significativas entre os grupos AD e HC a nível do voxel, o que sugere que a biodistribuição do PK no cérebro não permite diferenciar grupos. Esta afirmação apoia a associação que se tem vindo a estabelecer entre a microglia ativada e a neuroinflamação. Como a neuroinflamação é característica de cada indivíduo, isto é, é aleatoriamente distribuída no cérebro dos doentes de Alzheimer, o esperado é a não diferenciação de grupos por parte do PK. Foram encontradas cinco regiões cerebrais onde a correlação a nível do voxel se relevou elevada para quase todas as regiões de referência consideradas, córtex motor primário, córtex visual primário, córtex de associação somatossensorial, córtex visual associativo e córtex pré-motor. Tanto o precuneus (P) como o lóbulo parietal inferior (PI) desempenham funções importantes no processamento visual e espacial. Por conseguinte, é natural que os resultados da correlação a nível regional estejam associados com os obtidos a nível do voxel.Em suma, de acordo com o estudo realizado, o classificador construído apenas com features extraídas de imagens PET com PiB do grupo 4070, usando o cerebelo como região de referência, foi o classificador que apresentou uma melhor resposta ao problema proposto, classificação binária de indivíduos como AD ou HC. Adicionalmente, também foi descoberta uma correlação positiva entre o PK e o PiB em regiões cerebrais responsáveis pela função motora e pelo processamento visual.<br>Alzheimer’s disease (AD) is one of the main neurodegenerative disorders causing dementia. Positron emission tomography (PET) neuroimaging with 11C-Pittsburgh Compound B (PiB) and 11C-(R)-PK11195 (PK) are two of the existing modalities to assess amyloid plaque and activated microglia in human brain, respectively. Since amyloid plaque is the main hallmark of AD and activated microglia is currently found in the brain of AD patients, these imaging biomarkers can be used in diagnostic workup and to achieve early AD diagnosis.The main goal of the present study is to solve a binary classification problem between healthy controls (HC) and AD patients, by using machine learning (ML) methods based on two PET imaging biomarkers, PiB and PK. Another important goal of this work includes the identification of the brain regions where PiB and PK are most correlated, at both regional and voxel level.In the present study it was included 41 subjects (20 AD and 21 HC). To understand the impact of the time interval considered in PET image acquisition, the dataset was split in three different groups. Group TOT composed by PiB PET images acquired during the total time of PiB biodistribution, and groups 4070 and 4060, acquired during the characteristic accumulation time of PiB and PK, between minute 40 and 70, and 40 and 60, after administration, respectively. After quantification, pre-processing, feature extraction and selection, PiB and PK PET images were submitted to classification using a support vector machines (SVM) approach. Voxel-wise comparison between AD and HC groups of different quantified PK PET images were performed to understand the impact of distinct reference regions in the normalization of PK PET images and the influence of the quantification method used. Also, voxel-wise and region of interest (ROI) based correlation between standard uptake value ratio (SUVr) PiB and different quantified PK PET images were calculated.Normalization by cerebellum of PiB PET images of group 4070 yielded the best classification accuracy of AD (accuracy-0.925, sensitivity-1.000, specificity-0.857). Thus, for PiB PET images, cerebellum appears to be the brain region where amyloid accumulation bears the least significant differences between HC and AD patients, i.e., the best reference region to do the normalization. Also, when using the cerebellum as reference region of PiB PET images, stronger ROI-based correlation with binding potential (BP) PK PET images in several brain regions was found, compared to normalization based on white matter. Features extracted at regional level from PK PET images did not show improvement, neither in accuracy nor in sensitivity, of the classifier only based on features extracted from PiB PET images. ROI-based correlation results suggest specific binding of PK to cerebellum; thus, supervised cluster analysis algorithm based on four kinetic classes (SVCA4) showed to be the best approach to do the normalization of PK PET images. Both types of quantified PK PET images did not show significant differences between groups at voxel level. This suggests that PK biodistribution in the brain is not relevant for group differentiation. The reason why is probably related to the fact that activated microglia is associated with neuroinflammation, and this process is quite variable across participants, i.e., it is randomly distributed across brains of AD patients. There were five brain regions where the correlation at voxel level between PK and SUVr PiB PET images agreed the most for all reference regions considered, primary motor cortex, primary visual cortex, somatosensory association cortex, associative visual cortex and premotor cortex. Since, both precuneus (P) and parietal inferior (PI) have important roles in visuospatial processing, ROI-based correlation results are consistent with the ones obtained at voxel level.Overall, according with the present study, the classifier only based on features extracted from PiB PET images of group 4070, using cerebellum as reference region, was the classifier who solved more accurately the problem proposed, binary classification in AD. Additionally it was also found a positive correlation between PK and PiB in particular in brain regions responsible for motor function and visual processing.<br>FCT